This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: [3.3 branch] IA64 bootstrap failure
On Tue, Jul 08, 2003 at 02:54:24PM -0700, Zack Weinberg wrote:
> "H. J. Lu" <hjl@lucon.org> writes:
>
> > I found:
> >
> > /* If we don't have __builtin_constant_p, or it's not acceptable in
> > array initializers, fall back to assuming that all conditions
> > potentially vary at run time. It works in 3.0.1 and later; 3.0
> > only when not optimizing. */
> > #if (GCC_VERSION >= 3001) || ((GCC_VERSION == 3000) && !__OPTIMIZE__)
> > # define MAYBE_EVAL(expr) (__builtin_constant_p(expr) ? (int) (expr) : -1)
> > #else
> > # define MAYBE_EVAL(expr) -1
> > #endif
> >
> > in the generated insn-conditions.c. It means the stage1 compiler may
> > be different from the stage2 compiler, depending on what compiler is
> > used to bootstrap gcc. I think insn-conditions.c should be the same
> > regardless the compiler used to boostrap.
>
> Don't be ridiculous. The stage1 compiler is never going to be
> identical to the stage2 compiler. Ignoring for the moment the
> fact that stage1 is compiled with some random preexisting compiler
> with all optimizations off, there is a similar chunk of logic in the
> tree and RTL checking macros, which potentially affects every file in
> GCC.
>
That one is different. On ia64, there is INTEL_EXTENDED_IEEE_FORMAT,
which is defined as 1 for most platforms. Since ia64.md has 2 patterns
with !INTEL_EXTENDED_IEEE_FORMAT, as the result, the generated source
files are different, depending on what compiler is used to bootstrap.
H.J.